SPINE Volume 27, Number 4, pp 387–392
©2002, Lippincott Williams & Wilkins, Inc.
Adult Scoliosis
A Quantitative Radiographic and Clinical Analysis
Frank J. Schwab, MD,*† Vinson A. Smith, MD,* Michele Biserni, MD,‡ Lorenzo Gamez,§
Jean-Pierre C. Farcy, MD,*† and Murali Pagala, PhD储
Study Design. Prospective analysis of a consecutive
series of adult patients with adolescent idiopathic scolio-
sis of the adult and de novo degenerative scoliosis.
Objectives. To clinically and radiographically study
two populations of adult patients with either adolescent
idiopathic scoliosis of the adult or de novo degenerative
scoliosis in a quantitative manner to identify reliable radio-
graphic parameters that correlate with clinical symptoms.
Summary and Background. Although there are many
causes of spinal deformity in the adult, there are two main
categories of adult scoliosis: adolescent idiopathic scoli-
osis of the adult and de novo degenerative scoliosis. Un-
like pediatric scoliosis, in adults there are no established
radiographic parameters or classification systems that re-
liably provide a clinical correlation or offer a useful lan-
guage for communication among specialists. This study
gathered complete clinical and radiographic information
on 95 patients with adult scoliosis and established several
radiographic parameters that correlated with clinical
symptoms.
Methods. Each of the 95 patients completed a clinical
questionnaire that included a self-reported visual analog
scale and underwent full-length standing anteroposterior
and lateral radiography. Radiographic analysis was per-
formed by use of digital analysis and included measure-
ment of the Cobb angle, the number of vertebrae in each
curve, plumbline offset from T1 to the midsacral line, the
upper endplate obliquities of L3 and L4, and maximal
lateral olisthy between two adjacent lumbar vertebrae.
Sagittal plane measurements included lumbar lordosis,
thoracolumbar kyphosis, and the Sagittal Pelvic Tilt In-
dex. Statistical analysis of both radiographic and clinical
parameters of pain was performed to determine any sig-
nificant correlations between the two.
Results. This study showed that lateral vertebral olis-
thy, L3 and L4 endplate obliquity angles, lumbar lordosis,
and thoracolumbar kyphosis were significantly correlated
with pain.
Conclusion. This quantitative analysis identified sev-
eral clinically relevant radiographic parameters in adult
scoliosis patients. Additionally, excellent predictive for-
From the *Department of Orthopaedics, Maimonides Medical Center,
Brooklyn, New York, the †New York University School of Medicine,
New York, New York, the ‡Istituto di Scienze Ortopedico Trumato-
logiche e Riabilitative dell Universita di Siena, the §New York Medical
College, Valhalla, New York, and the 储Neuromuscular Research Lab,
Department of Surgery, Maimonides Medical Center, Brooklyn, New
York.
Supported by the Maimonides Medical Center Research and Develop-
ment Foundation.
Acknowledgment date: November 28, 2000.
First revision date: February 26, 2001.
Second revision date: April 5, 2001
Acceptance date: April 24, 2001.
Device status category: 1.
Conflict of interest category: 12.
mulas for self-reported pain levels were obtained. [Key
words: adult scoliosis, pain, radiographic analysis, lateral
olisthy] Spine 2002;27:387–392
Significant work has been dedicated to the diagnosis,
classification, and treatment of scoliosis in the pediatric
population. Based on radiographic parameters and nat-
ural history studies, classification systems and clinical
treatment guidelines have been established for idio-
pathic, congenital, neuromuscular, and metabolic scoli-
otic spinal deformities. By contrast, for adult scoliotic
deformities there are no clear diagnostic criteria, useful
classification systems, or accepted treatment guidelines.
A significant limitation in developing useful classifica-
tions and treatment guidelines for adult scoliosis has
been the limited understanding of relevant radiographic
parameters in these deformities.
The prevalence of adult scoliosis in the general popu-
lation has been reported as ranging from 1.4% to
12%.2,11,14 Aside from the esthetic considerations of this
physical deformity, significant pain and disability can
develop.1,3,4,12 With the demographic shift involving an
aging population in the United States and increased at-
tention to quality of life issues, adult scoliosis is becom-
ing a significant health care concern. The progression of
spinal deformities in the adult population, treatment ap-
proaches for adult scoliosis, and surgical techniques have
been reported in the literature.5,7,13,15–18
Adult scoliosis can be defined as a spinal deformity in
a skeletally mature patient with a Cobb angle greater
than 10°. Although there are many known causes of spi-
nal deformity in the adult, two categories include the
largest number of scolioses. The first category includes
patients with scoliosis during childhood and adolescence
that may progress or become symptomatic as the patient
ages. This type of scoliosis is often idiopathic and can be
termed adolescent scoliosis of the adult (ASA). The sec-
ond category includes patients in whom a the spinal de-
formity developed after skeletal maturity.6 This type of
scoliosis most commonly involves some form of degen-
erative disease and can be termed de novo degenerative
scoliosis (DDS). Although the causes of ASA and DDS
appear quite different, they may share a common path-
way in symptomatic patients: gradual loss of interseg-
mental stability with aging and consequent progressive
deformity and pain. Certainly, many adult deformities
may not fit clearly into the categories of ASA or DDS,
such as traumatic, metabolic, osteoporotic,16 or iatro-
genic deformities.
387
388 Spine • Volume 27 • Number 4 • 2002
Figure 1. Illustration of radiographic measurements taken from
standing spine x-ray.
The purpose of this study was to clinically and radio-
graphically study a population of adult patients with
ASA or DDS in a quantitative manner. The hypothesis
was that reliable radiographic parameters could be iden-
tified that would correlate with clinical symptoms in a
population of patients with adult scoliosis.
Materials and Methods
The study included a consecutive series of adult patients with
ASA or DDS in the practice of the two lead authors. The inclu-
sion criteria were as follows: patient age over 18 years, scoliosis
(thoracic, thoracolumbar, or lumbar) exceeding 15° by Cobb
angle, complete radiographic and clinical data. The exclusion
criteria were as follows: recent trauma, history of spinal sur-
gery, evidence of compression fracture or severe osteoporosis,
Figure 2. Illustration of radiographic measurements taken from
standing lateral spine radiographs.
Table 1. Correlation Between Radiographic Parameters
and Self-Reported Pain in Adult Scoliosis
Relevant Nonrelevant
Olisthy Cobb angle
L3 Endplate angle Age
L4 Endplate angle Level of olisthy
Lumbar lordosis SPTI
Thoracolumbar kyphosis Plumbline offset
SPTI ⫽ Sagittal Pelvic Tilt Index.
metabolic bone disease, or underlying pathologic state of the
bone such as metastatic disease.
A series of 95 patients met the above criteria. Each patient
completed a detailed questionnaire, which reviewed demo-
graphic data, medical history, and type, location, and duration
of pain. The self-reported Visual Analog Scale (VAS) was ap-
plied to evaluate the severity of pain. In addition to the clinical
questionnaire, each patient underwent a complete radiographic
series, including full-length standing anteroposterior and lat-
eral radiographs (scoliosis series).
Radiographic Analysis. The full-length spine radiographs
were evaluated for several parameters. All measurements were
made by digital analysis, using the Xcalliper (Eisenlohr Corp.,
Davis, CA). From the frontal plane (anteroposterior radio-
graphs), the following were recorded: Cobb angle of all scoli-
otic curvatures (thoracic, thoracolumbar, lumbar), number of
vertebrae in each curve, plumbline offset from T1 to midsacral
line, endplate obliquities (from the horizontal) of L3 and L4
vertebrae, and maximal lateral olisthy (offset) between any two
adjacent lumbar vertebrae (Figure 1). From the sagittal plane
(lateral radiographs) the following were recorded: lumbar lor-
dosis (upper endplate L1 vertebra to upper endplate S1 verte-
bra, Cobb angle), thoracolumbar kyphosis (upper endplate
T11 vertebra to lower endplate L1 vertebra, Cobb angle), sag-
ittal pelvic tilt index (Figure 2).
Figure 3. Relation between Visual Analog Scale (VAS) and olisthy.

Figure 4. Relation between Visual Analog Score (VAS) and L3
endplate obliquity.
Statistical Analysis. The clinical parameters of pain (leg or
back), intensity of pain by self-reported VAS, and radiographic
parameters were entered into a SigmaStat (SPSS Corp., Chi-
cago, IL) worksheet. A statistician not involved with data ac-
quisition and blinded to patient identity evaluated the correla-
tions between pain and radiographic parameters. Linear
regression analysis was applied to assess the relation between
the VAS score of pain and each of the relevant radiographic
parameters to obtain predictive formulas.
Results
The study population included 95 patients, 33 men and
62 women. The average age was 59 years (range, 18 – 88
Figure 5. Relation between Visual Analog Score (VAS) and L4
endplate obliquity.
Quantitative Analysis of Adult Scoliosis • Schwab et al 389
Figure 6. Relation between Visual Analog Score (VAS) and lumbar
lordosis.
years). Some degree of pain was reported in 78 patients
(74%). The mean VAS score for the symptomatic pa-
tients in the study group was 58 (range, 20 –100). The
type of scoliotic deformity was DDS in 57 patients (54%)
and ADA in 38 patients (46%). Scoliotic curvatures of
the lumbar or thoracolumbar spine averaged 36° (SD,
18°; range, 15–72°), and curvatures of the thoracic spine
averaged 28° (SD, 22°; range, 10 –120°).
After statistical analysis of both radiographic and
clinical data, significant correlations between radio-
Figure 7. Relation between Visual Analog Score (VAS) and thora-
columbar kyophosis.
390 Spine • Volume 27 • Number 4 • 2002

Table 2. Predictive Formulas for Self-Reported Pain in L3 endplate obliquity angle was significant (r ⫽ 0.433,
Study Group of Adult Scoliosis Patients: Visual Analog P ⬍ 0.002). The predictive formula was VAS ⫽ 45.96 ⫹
Pain Score and Radiographic Parameters (0.833 ⫻ L3 angle) (Figure 4). The VAS and L4 endplate
angle showed a significant correlation (r ⫽ 0.375, P ⬍
Level of
Formula Significance (P) 0.006). The predictive formula was VAS ⫽ 42.95 ⫹
(0.934 ⫻ L4 angle) (Figure 5). In terms of sagittal plane
VAS ⫽ 42.947 ⫹ (0.934 ⫻ L4 angle) ⬍0.006 radiographic analysis, the correlation between VAS and
VAS ⫽ 45.96 ⫹ (0.833 ⫹ L3 angle) ⬍0.002
VAS ⫽ 41.52 ⫹ (3.152 ⫻ olisthy) ⬍0.001 lumbar lordosis was significant (r ⫽ 0.465, P ⬍ 0.001).
VAS ⫽ 86.353 ⫺ (0.695 ⫻ lordosis) ⬍0.001 The predictive formula was VAS ⫽ 86.353 ⫺ (0.695 ⫻
VAS ⫽ 39.612 ⫹ (1.720 ⫻ kyphosis) ⬍0.001 lordosis) (Figure 6). The correlation between VAS and
VAS ⫽ Visual Analog Scale. thoracolumbar kyphosis was significant (r ⫽ 0.508, P ⬍
0.001. The predictive formula was VAS ⫽ 39.612 ⫹
(1.720 ⫻ kyphosis) (Figure 7).
graphic and clinical parameters (pain) were identified. There was no statistically significant correlation be-
From standing anteroposterior radiographs it was deter- tween the VAS and the number of vertebrae involved in
mined that lateral vertebral olisthy, and L3 and L4 end- the scoliotic Cobb angle. Even more convincing was that
plate obliquity angles, were correlated with VAS pain the Cobb angle of the scoliotic deformity had no statis-
score. From standing lateral radiographs it was deter- tically significant correlation to the VAS. The sagittal
mined that lumbar lordosis and thoracolumbar kyphosis pelvic tilt index and plumbline offset showed no statisti-
were statistically significantly correlated with self- cally significant correlation with self-reported pain by
reported pain (Table 1). VAS (Table 2).
In terms of frontal plane radiographic analysis, the
correlation between VAS and lateral olisthy was statisti-
Discussion
cally significant (r ⫽ 0.473, P ⬍ 0.001). The predictive
formula was VAS ⫽ 41.52 ⫹ (3.152 ⫻ olisthy) (Figure Adult scoliotic deformities have received little attention.
3). The level of lateral vertebral olisthy from L1 to L5 did In part this may result from the impression of limited
not bear a statistically significant correlation with the scoliotic progression in this group and the focus of treat-
VAS. In other words, although the amount of olisthy was ment on regional degenerative pathologic conditions
highly significant, the level of the olisthy had no impact (stenosis, spondylolisthesis, disc degeneration) rather
on patient pain scores. The correlation between VAS and than the deformity itself. An additional reason may lie in

Figure 8. A, Standing anteropos-

terior radiograph of the spine in a
subject with minimal back dis-
comfort, showing a marked sco-
liotic deformity. The patient re-
ceived a diagnosis of idiopathic
scoliosis as an adolescent. B,
Standing lateral radiograph of
the spine in the same patient. A
well-preserved lumbar lordosis is
evident.

Figure 9. A, Standing anteroposterior radiograph of the spine in a patient with significant back pain. Scoliosis was not diagnosed during
adolescence. B, Anteroposterior lumbar radiograph, showing significant lateral olisthy at L3–L4. C, Standing lateral radiograph of the
patient. A marked loss of lumbar lordosis is evident.
the limited surgical treatment in the elderly who have
deformities frequently associated with other medical
conditions. For several reasons the perceptions of scoli-
osis in the adult may be changing. Notable demographic
shifts in Western societies and an increased life expect-
ancy are evident, and quality of life concerns are focusing
attention on the degenerative pathologic conditions in
aging persons who wish not only to age but to do so with
independence in daily living.
Although the common degenerative conditions of the
spine are frequently treated as focal pathologic states, it
appears intuitive that deformity of the spinal column, by
altering the mechanical loading conditions, can acceler-
ate the degenerative cascade. The question that arises,
therefore, is whether radiographic parameters can be
identified in an adult with scoliosis that will serve a pre-
dictive purpose on a clinical level.8 –10 Furthermore, it
appears that radiographic criteria should be developed
for the adult scolioses to serve as a basis of analysis in
longitudinal studies and perhaps eventually for treat-
ment guidelines.
The findings in this study clearly illustrate that tradi-
tional parameters such as Cobb angle and age are not
useful in the assessment of adult scoliosis ( Figures 8 and
9). The Cobb angles of patients in this study covered a
wide range, but this parameter itself was not a significant
criterion in this group of adult scoliosis patients, al-
Quantitative Analysis of Adult Scoliosis • Schwab et al 391
though its role in pediatric deformity is paramount. Pa-
tients with ASA had larger major scoliosis curvatures
than patients with DDS (ASA mean, 40°; DDS mean,
25°; although the radiographic parameters correlating
with pain were identical for these groups). This appears
to substantiate the belief that a common end pathway
(degenerative instability and unfavorable lumbar verte-
bral alignment) among both groups of patients is related
to symptoms rather than the degree of curvature or the
cause of the original scoliosis.
In this quantitative analysis, several clinically relevant
radiographic criteria in two adult scoliosis populations
were identified. Excellent predictive formulas for self-
reported pain levels on the VAS were obtained. The most
significant radiographic parameters in the adult scoliotic
group were upper endplate obliquities of L3 and L4,
lateral olisthy between adjacent lumbar vertebrae on the
frontal plane, lumbar lordosis, and thoracolumbar
kyphosis.
The highly significant radiographic parameters iden-
tified in this study population of adult scoliosis patients
appear to reflect the level of regional balance, instability,
and pathologic mechanical loads of the spinal elements.
It is not clear to what degree the individual parameters
are interrelated and whether progression beyond a cer-
tain level of intervertebral alignment of one factor will
inevitably lead to changes in other measured parameters.
392 Spine • Volume 27 • Number 4 • 2002
That is, we cannot yet conclude that a certain degree of
endplate obliquity in the frontal plane will lead to inev-
itable vertebral olisthy over time and thence to changes
in sagittal plane contour and to pain. Such a theory
would be appealing, and clearly longitudinal studies are
important in this area. Once such a link can be con-
firmed, treatment guidelines may be strongly affected,
and early intervention in a middle-aged adult with scoli-
osis may, in some cases, be preferable to treating ad-
vanced deformity in that same person once he or she has
become elderly.
Statistically highly significant radiographic parame-
ters in the setting of clinical symptoms and adult scoliosis
have been identified. These data are encouraging, and the
establishment of a multicenter database is anticipated.
Further investigation will permit increased statistical
power of these findings and may lay the groundwork for
a classification system of the adult scolioses.
Key Points
● There is a correlation between radiographic pa-
rameters and pain in adult scoliosis.
● Highly significant radiographic parameters in-
clude endplate obliquities of L3 and L4, lateral olis-
thy between the lumbar vertebrae, lumbar lordosis,
and thoracolumbar kyphosis.
Acknowledgments
The authors thank Sobhan Pagala for statistical support.
References
1. Balderston RA. Adult scoliosis: The thoracic spine. Chapter 46 in The
Textbook of Spinal Surgery, 2nd ed. Philadelphia: Lippincott-Raven,
1977:715–32.
2. Biot B, Pendrix D. Frequence de la scoliose lombaire a l’age adulte. Ann Med
Phys 1982;25:251– 4.
3. Bradford D, Tay B, Hu S. Adult scoliosis: surgical indications, operative
management, complications, and outcomes. Spine 1999;24:2617–29.
4. Bradford D. Adult scoliosis: current concepts of treatment. Clin Orthop
1988;229:70 – 87.
5. Duval-Beaupere G, Lamireau TH. Scoliosis at less than 30 degrees: Proper-
ties of the evolutivity (risk of progression). Spine 1985;10:421– 4.
6. Grubb S, Libscomb H, Coonrad R. Degenerative adult-onset scoliosis. Spine
1988;13:241–5.
7. Grubb S, Lipscomb H, Suh P. Results of surgical treatment of painful adult
scoliosis. Spine 1994;19:1619 –27.
8. Kopec JA. Measuring functional outcomes in persons with back pain: A
review of back-specific questionnaires. Spine 2000;25:3110 – 4.
9. Jackson R, Peterson M, McManus A, et al. Compensatory spinopelvic bal-
ance over the hip axis and better reliability in measuring lordosis to the pelvic
radius on standing lateral radiographs of adult volunteers and patients. Spine
1998;16:1750 – 67.
10. Jackson R, Simmons E, Stripinis D. Coronal and sagittal plane spinal defor-
mities correlating with back pain and pulmonary function in adult idiopathic
scoliosis. Spine 1989;14:1391–7.
11. Korovessis P, Piperos G, Sidiropoulos P, et al. Adult idiopathic lumbar sco-
liosis: A formula for prediction of progression and review of the literature.
Spine 1994;19:1926 –32.
12. Kostuik JP, Bentivoglio J. The incidence of low-back pain in adult scoliosis.
Spine 1981;6:268 –73.
13. Kostuik JP. Adult scoliosis: The lumbar spine. Chapter 47 in The Textbook
of Spinal Surgery, 2nd ed. Philadelphia: Lippincott-Raven, 1977:733–75.
14. Perdriolle R, Vidal J. Thoracic idiopathic scoliosis curve evolution and prog-
nosis. Spine 1985;10:785–91.
15. Perennou D, Marcelli C, Herisson C. Adult lumbar scoliosis: Epidemiologic
aspects in a low-back pain population. Spine 1994;19:123– 8.
16. Thevenon A, Pollez B, Cantegrit F, et al. Relationship between kyphosis,
scoliosis, and osteoporosis in the elderly population. Spine 1987;12:744 –5.
17. Velis KP, Healey JH, Schneider R. Osteoporosis in unstable adult scoliosis.
Clin Orthop 1988;237:132– 41.
18. Yasuo Y, Yamaguchi T, Asaka Y. Prediction of curve progression in idio-
pathic scoliosis based on initial roentgenograms: A proposal of an equation.
Spine 1988;13:1258 – 61.
Address reprint requests to
Frank J. Schwab, MD
927 49th Street
Brooklyn, NY 11219
E-mail: spinecenter@orthospine.com